The exponential proliferation of data in a multitude of domains such as business, organizational systems,health and finance has necessitated the emergence of intelligent analytical frameworks capable of supporting the process of decision making based on extensive data sets. Isolated tools coupled with manual data analysis often hinder effective decision-making in large and complex data sets. In light of these challenges, this dissertation advocates for the construction of a holistic visual analytics framework combining the functions of data pre-processing, machine learning, temporal analysis, and visualization into one system. This system allows users to upload data sets, process the data end-to-end with the aid of an intuitive interface, and execute various analysis methods. These methods span across various domains, from selection of relevant data features, anomaly detection and data clustering to temporal monitoring and automatic extraction of insights.We employ methods such as K-Means clustering, Isolation Forest and regression models in order to extract patterns, analyze time dependent variations, anomalies, and build predictive models respectively. Time based analysis provides valuable insights about changes in time and helps to identify trends and patterns.We provide a unique feature whereby the system is capable of generating automatic insights based on the analytical output, thus helping the users who have limited technical skills to understand and use the information at hand. It also facilitates handling and comparison of multiple data sets and analysis of differences among data sets and performance variations.The proposed framework is modular which aids it to remain flexible, scalable and adaptable. We aim to make it generic in nature so it can be used with all forms of data including numerical, categorical and temporal attributes and various domains of applications.Experimental evaluation of the prototype has proven that it effectively identifies trends, anomalies and extract the necessary patterns, enabling informed decision making through automated insights and effective visualization. This project successfully illustrates the construction of an efficient visual analytics tool in aid of an effective decision support system.

Key Words: Visual Analytics, Decision Support System, Machine Learning, Temporal